Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> How much did you bloat the code? There are an awful lot of calls to
> newNode(), so even though it's not all that large, I'd think the
> multiplier would be nasty.
The patch increases the executable from 12844452 to 13005244 bytes,
when compiled with '-pg -g -O2' and without being stripped.
> This isn't portable at all, AFAIK :-(. Unfortunately I can't think
> of a portable way to do it with a macro, either.
Well, one alternative might be to provide 2 definitions of the
function -- one an extern inline in the header file, and one using the
current method (in a separate file, non-inline). Then wrap the header
file in an #ifdef __GNUC__ block, and the non-inlined version in
#ifndef __GNUC__. The downside is that it means maintaining two
versions of the same function -- but given that newNode() is pretty
trivial, that might be acceptable.
BTW, the GCC docs on inline functions are here:
According to that page, using 'static inline' instead of 'extern
inline' is recommended for future compatability with C99, so that's
what we should probably use (in the __GNUC__ version).
Neil Conway <neilc(at)samurai(dot)com> || PGP Key ID: DB3C29FC
In response to
pgsql-hackers by date
|Next:||From: Curtis Faith||Date: 2002-10-07 23:04:58|
|Subject: Re: Analysis of ganged WAL writes|
|Previous:||From: Alvaro Herrera||Date: 2002-10-07 21:45:57|
|Subject: Re: 7.2.3 patching done|
pgsql-patches by date
|Next:||From: Tom Lane||Date: 2002-10-08 01:08:20|
|Subject: Re: inline newNode() |
|Previous:||From: Kris Jurka||Date: 2002-10-07 21:58:51|
|Subject: Re: DBMD Patch|